Implementation of a brain-computer interface based on three states of motor imagery

A motor imagery based brain-computer interface (BCI) translates the subject's motor intention into a control signal through real-time detection of characteristic EEG spatial distributions corresponding to motor imagination of different body parts. In this paper, we implemented a three-class BCI manipulated through imagination of left hand, right hand and foot movements, inducing different spatial patterns of event-related desynchronization (ERD) on mu rhythms over the sensory-motor cortex. A two-step training approach was proposed including consecutive steps of online adaptive training and offline training. Then, the optimized parameters and classifiers were utilized for online control. This paradigm facilitated three directional movement controls which could be easily applied to help the motion-disabled to operate a wheelchair. The average online and offline classification accuracy on five subjects was 79.48% and 85.00% respectively, promoting the three-class motor imagery based BCI a promising means to realize brain control of a mobile device.